Resource recommendation method and device and electronic equipment

Abstract

The invention provides a resource recommendation method and device and electronic equipment, and relates to the technical field of artificial intelligence, in particular to the technical fields of deep learning, cloud computing, natural language processing, intelligent recommendation, large models and the like. According to the specific implementation scheme, a candidate resource set and a historical resource sequence of a target object in a historical time period before a current time point are obtained; determining an interest representation vector of the target object at the current time point according to the historical resource sequence; the interest representation vector represents resources interested by the target object at the current time point; selecting to-be-recommended resources from the candidate resource set according to the interest representation vector, and further performing resource recommendation processing; wherein the interest representation vector can represent the resource interested by the target object at the current time point, so that the determined to-be-recommended resource is the resource which the target object wants to click at present, the matching degree between the to-be-recommended resource and the content interested by the target object is improved, and the resource recommendation accuracy and the resource recommendation efficiency are improved.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to technical fields such as deep learning, cloud computing, natural language processing, intelligent recommendation, and large models, and in particular to a resource recommendation method, device, and electronic device. Background Art

[0002] In the current resource recommendation method, the characteristics of the object are learned based on the historical consumption resources of the object; based on the characteristics of the object, resource recall processing is performed to obtain the resources to be recommended, and resource recommendation processing is performed.

[0003] Among them, in the above scheme, the characteristics of the object represent the object's long-term resource preference, which may not match the resource that the object currently wants to click, thereby reducing the accuracy and efficiency of resource recommendation. Summary of the invention

[0004] The present disclosure provides a resource recommendation method, device and electronic device.

[0005] According to one aspect of the present disclosure, a resource recommendation method is provided, the method comprising: obtaining a candidate resource set and a historical resource sequence of a target object in a historical time period before a current time point; determining an interest representation vector of the target object at the current time point based on the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point; selecting a resource to be recommended from the candidate resource set based on the interest representation vector; and performing resource recommendation processing on the target object based on the resource to be recommended.

[0006] According to another aspect of the present disclosure, a method for training a resource recommendation model is provided, the method comprising: obtaining training data; the training data comprising structured content of sample historical resources in a sample historical resource sequence, structured content of sample recommendation resources, and sample click resources in the sample recommendation resources; obtaining an initial resource recommendation model; the resource recommendation model comprising an encoding network, a decoding network, and a fully connected network connected in sequence; training the initial resource recommendation model using the structured content of the sample historical resources, the structured content of the sample recommendation resources, and the sample click resources to obtain a trained resource recommendation model.

[0007] According to another aspect of the present disclosure, a resource recommendation device is provided, the device comprising: an acquisition module for acquiring a candidate resource set and a historical resource sequence of a target object in a historical time period before a current time point; a determination module for determining an interest representation vector of the target object at the current time point based on the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point; a selection module for selecting a resource to be recommended from the candidate resource set based on the interest representation vector; and a recommendation processing module for performing resource recommendation processing on the target object based on the resource to be recommended.

[0008] According to another aspect of the present disclosure, a training device for a resource recommendation model is provided, the device comprising: a first acquisition module, used to acquire training data; the training data comprises structured content of sample historical resources in a sample historical resource sequence, structured content of sample recommendation resources, and sample click resources in the sample recommendation resources; a second acquisition module, used to acquire an initial resource recommendation model; the resource recommendation model comprises an encoding network, a decoding network, and a fully connected network connected in sequence; a training processing module, used to perform training processing on the initial resource recommendation model using the structured content of the sample historical resources, the structured content of the sample recommendation resources, and the sample click resources to obtain a trained resource recommendation model.

[0009] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the resource recommendation method proposed above in the present disclosure; or, execute the training method of the resource recommendation model proposed above in the present disclosure.

[0010] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the resource recommendation method proposed above in the present disclosure; or, to execute the training method of the resource recommendation model proposed above in the present disclosure.

[0011] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps of the resource recommendation method proposed above in the present disclosure; or, implements the steps of the training method of the resource recommendation model proposed above in the present disclosure.

[0012] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0014] Figure 1 is a schematic diagram according to a first embodiment of the present disclosure;

[0015] Figure 2 is a schematic diagram according to a second embodiment of the present disclosure;

[0016] Figure 3 is a schematic diagram according to a third embodiment of the present disclosure;

[0017] Figure 4 is a schematic diagram according to a fourth embodiment of the present disclosure;

[0018] Figure 5 is a schematic diagram according to a fifth embodiment of the present disclosure;

[0019] Figure 6 is a schematic diagram according to a sixth embodiment of the present disclosure;

[0020] Figure 7 It is a block diagram of an electronic device used to implement the resource recommendation method or the resource recommendation model training method of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0021] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0022] In the current resource recommendation method, the characteristics of the object are learned based on the historical consumption resources of the object; based on the characteristics of the object, resource recall processing is performed to obtain the resources to be recommended, and resource recommendation processing is performed.

[0023] Among them, in the above scheme, the characteristics of the object represent the object's long-term resource preference, which may not match the resource that the object currently wants to click, thereby reducing the accuracy and efficiency of resource recommendation.

[0024] In view of the above problems, the present disclosure proposes a resource recommendation method, device and electronic device.

[0025] Figure 1 It is a schematic diagram according to the first embodiment of the present disclosure. It should be noted that the resource recommendation method of the embodiment of the present disclosure can be applied to a resource recommendation device, which can be configured in an electronic device so that the electronic device can perform a resource recommendation function.

[0026] Among them, the electronic device can be any device with computing capabilities, such as a personal computer (PC), a mobile terminal, a server, etc. The mobile terminal can be, for example, a vehicle-mounted device, a mobile phone, a tablet computer, a personal digital assistant, a wearable device, a smart speaker, a server, a server cluster, and other hardware devices with various operating systems, touch screens and / or display screens.

[0027] The resource recommendation device may also be software in an electronic device, such as resource recommendation software, etc. In the following embodiments, the execution subject is an electronic device as an example for description.

[0028] like Figure 1 As shown, the resource recommendation method may include the following steps:

[0029] Step 101: Acquire a candidate resource set and a historical resource sequence of a target object in a historical time period before a current time point.

[0030] In the disclosed embodiment, the process of the electronic device executing step 101 may, for example, be to perform resource recall processing and resource screening processing according to the attribute information of the target object to obtain a set of candidate resources; obtain the historical resources clicked by the target object in the historical time period before the current time point, and the click time points of the historical resources; and sort the historical resources according to the click time points to obtain a historical resource sequence.

[0031] Among them, the historical resources clicked by the target object in the historical time period before the current time point are the target object's behavioral data on resources, which can represent the resources that the target object is interested in in the historical time period; and then combined with the historical resources, it is possible to accurately predict the resources that the target object may be interested in at the current time point, thereby improving the accuracy of the determined interest representation vector, and improving the match between the determined resources to be recommended and the content of interest to the target object.

[0032] The attribute information of the target object may include at least one of the following: geographic location information, attribute information of the device used, time information, age, gender, occupation type, etc. The attribute information of the device used includes at least one of the following: usage records of each application in the device, device usage time, location information when the device is used, etc., which can be set according to actual needs and are not specifically limited here.

[0033] The acquisition and use of attribute information of user objects in the embodiments of the present disclosure are known and agreed to by users, are in compliance with relevant laws and regulations, and do not violate public order and good morals.

[0034] In an embodiment of the present disclosure, the process of an electronic device determining a candidate resource set based on the attribute information of a target object may, for example, be: performing resource recall processing based on the attribute information of the target object to obtain a recalled resource set; determining a first click rate of recalled resources in the recalled resource set based on the attribute information and a first click rate model; performing screening processing on recalled resources in the recalled resource set based on the first click rate to obtain a filtered resource set; determining a second click rate of filtered resources in the filtered resource set based on the attribute information and a second click rate model; the parameter amount of the second click rate model is greater than the parameter amount of the first click rate model; and performing screening processing on filtered resources in the filtered resource set based on the second click rate to obtain a candidate resource set.

[0035] The electronic device may input the attribute information of the target object and the recalled resources in the recalled resource set into a first click rate model to obtain a first click rate output by the first click rate model. The electronic device may input the attribute information of the target object and the filtered resources in the filtered resource set into a second click rate model to obtain a second click rate output by the second click rate model.

[0036] Among them, the electronic device's screening process for recalled resources based on the first click rate of recalled resources in the recalled resource set can, for example, be: for each recalled resource, determine whether the first click rate of the recalled resource is less than a first click rate threshold; if the first click rate is less than the first click rate threshold, filter the recalled resource; if the first click rate is greater than or equal to the first click rate threshold, retain the recalled resource.

[0037] The electronic device may screen the screening resources according to the second click rate of the screening resources in the screening resource set, for example, by determining, for each screening resource, whether the second click rate of the screening resource is less than a second click rate threshold; filtering the screening resource if the second click rate is less than the second click rate threshold; and retaining the screening resource if the second click rate is greater than or equal to the second click rate threshold. The first click rate threshold may be less than the second click rate threshold.

[0038] The parameter amount of the second click rate model is greater than that of the first click rate model. By combining the first click rate model with the recalled resources in the recalled resource set, the number of filtered resources that the second click rate model needs to process can be reduced, thereby reducing the processing amount of the second click rate model and improving processing efficiency.

[0039] Step 102, determining an interest representation vector of the target object at the current time point according to the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point.

[0040] In the disclosed embodiment, the interest representation vector may be a representation vector of a resource that the target object is interested in at the current time point. Correspondingly, the electronic device may determine the interest representation vector in combination with the interest prediction model, and further determine the resource to be recommended to the target object. The interest prediction model may be trained in combination with the historical resource sequence of the target object.

[0041] It is assumed that the historical resource sequence includes N historical resources. The training process of the interest prediction model can be, for example, inputting the first N-1 historical resources into the interest prediction model to obtain the predicted interest representation vector output by the interest prediction model; combining the predicted interest representation vector to determine the Nth predicted resource; combining the Nth predicted resource and the Nth historical resource to determine the value of the loss function of the interest prediction model; training the interest prediction model according to the value to obtain the trained interest prediction model.

[0042] Step 103: Select a resource to be recommended from the candidate resource set according to the interest representation vector.

[0043] In an embodiment of the present disclosure, in one example, the electronic device may determine the similarity between the interest representation vector and the representation vectors of candidate resources in the candidate resource set; and select a resource to be recommended from the candidate resource set based on the similarity.

[0044] Step 104: perform resource recommendation processing on the target object according to the resources to be recommended.

[0045] The resource recommendation method of the embodiment of the present disclosure obtains a candidate resource set and a historical resource sequence of the target object in a historical time period before the current time point; determines an interest representation vector of the target object at the current time point based on the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point; selects a to-be-recommended resource from the candidate resource set based on the interest representation vector; and performs resource recommendation processing on the target object based on the to-be-recommended resource; wherein the interest representation vector determined in combination with the historical resource sequence can represent the resources that the target object is interested in at the current time point, so that the determined to-be-recommended resource is the resource that the target object currently wants to click, thereby improving the matching degree between the to-be-recommended resource and the content of interest to the target object, and improving the accuracy of resource recommendation; and accurate recommendation of resources can reduce the number of resource recommendation times, thereby improving resource recommendation efficiency.

[0046] In order to improve the accuracy of the interest representation vector determined, more features of the historical resources in the historical resource sequence can be considered. Correspondingly, the electronic device can first determine the representation vector of the historical resources in the historical resource sequence, and then determine the interest representation vector of the target object in combination with the representation vector of the historical resources. Figure 2 As shown, Figure 2 is a schematic diagram according to a second embodiment of the present disclosure, Figure 2 The illustrated embodiment may include the following steps:

[0047] Step 201: Acquire a candidate resource set and a historical resource sequence of a target object in a historical time period before a current time point.

[0048] Step 202: Determine the characterization vector of the historical resource in the historical resource sequence.

[0049] In the embodiment of the present disclosure, the process of the electronic device executing step 202 may, for example, be to determine the structured content of the historical resources in the historical resource sequence; the structured content includes at least one metadata in the historical resources; the structured content is input into the encoding network in the resource recommendation model, and the representation vector of the historical resources output by the encoding network is obtained.

[0050] The metadata in the structured content may include at least one of the following: the title of the resource, the keywords in the resource, the domain to which the resource belongs, and the publication object of the resource. The domain to which the resource belongs may be the domain of the resource at various domain levels. For example, the domain to which the resource belongs may include at least one of the following: the first-level domain to which the resource belongs, the second-level domain to which the resource belongs under the first-level domain, the third-level domain to which the resource belongs under the second-level domain, etc., which are not specifically limited here and can be set according to actual needs.

[0051] Among them, as an alternative solution, the electronic device can perform word segmentation processing on the metadata in the structured content to obtain individual words; determine the representation vector of each word; splice the representation vector of each word to obtain a first splicing processing result; and use the first splicing processing result as the structured content to be input into the encoding network.

[0052] As another alternative, the electronic device can set a number for each word obtained by the segmentation process; store the corresponding relationship between the representation vector and the code of each word. The numbers of each word are spliced ​​to obtain a second splicing result; the second splicing result is used as the structured content to be input into the encoding network. Correspondingly, the encoding network can query the representation vector corresponding to each number in the second splicing result; and then obtain the representation vector of each word; according to the representation vector of each word, the representation vector of the resource is determined.

[0053] Among them, the setting of structured content makes it possible for the encoding network to determine the representation vector of the newly released resources in combination with the structured content of the newly released resources without the need for learning processing, so that the resource recommendation method disclosed in the present invention can be applied to recommend newly released resources, thereby further improving the efficiency of resource recommendation.

[0054] Step 203: Determine the interest representation vector of the target object at the current time point according to the representation vector of each historical resource in the historical resource sequence.

[0055] In the embodiment of the present disclosure, the process of the electronic device executing step 203 may, for example, be to determine a representation vector sequence based on the representation vectors of each historical resource in the historical resource sequence; input the representation vector sequence into a decoding network in the resource recommendation model, and obtain the interest representation vector output by the decoding network.

[0056] Among them, the interest representation vector is determined by combining the representation vector sequence and the decoding network, so that the decoding network can learn the correlation between each representation vector in the representation vector sequence, and learn the changing trend of the resources that the target object is interested in, thereby improving the accuracy of the interest representation vector of the target object at the current time point, and further improving the efficiency of resource recommendation.

[0057] In the disclosed embodiment, based on the fact that the resources of interest to the target object will not change in a short period of time, in order to weaken the influence of the order of historical resources in the historical resource sequence on the interest representation vector, it is possible to consider dividing the historical resources in the historical resource sequence according to the time point to obtain multiple historical resource combinations; and determine the interest representation vector in combination with the combination to which the historical resources in the historical resource sequence belong. Correspondingly, before step 203, the electronic device may also perform the following process: divide and process each historical resource in the historical resource sequence according to the acquisition time point to obtain multiple historical resource combinations; determine the combination number of the historical resource combination, and the number vector corresponding to the combination number; for each historical resource in the historical resource sequence, concatenate the number vector corresponding to the combination number of the historical resource combination to which the historical resource belongs with the representation vector of the historical resource to obtain a processed representation vector; and update the representation vector of the historical resource according to the processed representation vector.

[0058] Among them, the electronic device can divide the time period according to the collection time point of each historical resource in the historical resource sequence to obtain multiple historical time periods; each historical time period corresponds to a historical resource combination; the collection time point of the historical resources in the historical resource combination is located in the historical time period corresponding to the historical resource combination; for each historical resource combination, a combination number is set; the combination number is vectorized to obtain a number vector.

[0059] Among them, historical resources belonging to the same historical resource combination are set with the same combination number, so that the time sequence between the historical resources can be weakened to a certain extent, reducing the influence of the time sequence between the historical resources on the interest representation vector.

[0060] Step 204: Select a resource to be recommended from the candidate resource set according to the interest representation vector.

[0061] Step 205: Perform resource recommendation processing on the target object according to the resources to be recommended.

[0062] It should be noted that the details of step 201, step 204 to step 205 can be found in Figure 1 Step 101, step 103 to step 104 in the illustrated embodiment will not be described in detail here.

[0063] The resource recommendation method of the embodiment of the present disclosure obtains a candidate resource set and a historical resource sequence of the target object in a historical time period before the current time point; determines the representation vectors of the historical resources in the historical resource sequence; determines the interest representation vector of the target object at the current time point based on the representation vectors of each historical resource in the historical resource sequence; selects a to-be-recommended resource from the candidate resource set based on the interest representation vector; and performs resource recommendation processing on the target object based on the to-be-recommended resource; wherein, by first determining the representation vectors of the historical resources in the historical resource sequence and then determining the interest representation vector of the target object based on the representation vectors of the historical resources, more features of the historical resources can be considered, and further ensuring that the determined to-be-recommended resource is the resource that the target object currently wants to click, thereby improving the matching degree between the to-be-recommended resource and the content of interest of the target object, and improving the accuracy of resource recommendation; and accurate recommendation of resources can reduce the number of resource recommendations, thereby improving resource recommendation efficiency.

[0064] In order to further improve the matching degree between the determined resources to be recommended and the target object, and to improve the click rate of the target object on the recommended resources, the predicted click rate of the candidate resources can be determined based on the interest representation vector and the representation vector of the candidate resources in the candidate resource set, and then the resources to be recommended can be selected based on the predicted click rate. Figure 3 As shown, Figure 3 is a schematic diagram according to a third embodiment of the present disclosure, Figure 3 The illustrated embodiment may include the following steps:

[0065] Step 301: Acquire a candidate resource set and a historical resource sequence of a target object in a historical time period before a current time point.

[0066] Step 302: Determine an interest representation vector of the target object at the current time point based on the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point.

[0067] Step 303: Determine the representation vector of the candidate resource in the candidate resource set.

[0068] In the embodiment of the present disclosure, the process of the electronic device executing step 303 may, for example, be to determine the structured content of the candidate resource; the structured content includes at least one metadata in the candidate resource; the structured content of the candidate resource is input into the encoding network in the resource recommendation model, and the representation vector of the candidate resource output by the encoding network is obtained.

[0069] Step 304: Determine the predicted click rate of the candidate resource according to the interest representation vector and the representation vector of the candidate resource.

[0070] In the embodiment of the present disclosure, the process of the electronic device executing step 304 may be, for example, inputting the interest representation vector and the representation vector of the candidate resource into a fully connected network in the resource recommendation model, and obtaining the predicted click rate output by the fully connected network.

[0071] Among them, the resource recommendation model can be a trained resource recommendation model, which is combined with the fully connected network in the resource recommendation model to determine the predicted click rate of the candidate resources, which can improve the accuracy of the determined predicted click rate.

[0072] Step 305: Select a resource to be recommended from the candidate resource set according to the predicted click rate.

[0073] In an embodiment of the present disclosure, in one example, the process of the electronic device executing step 305 may be, for example, to sort the candidate resources in the candidate resource set in descending order according to the predicted click-through rate to obtain a sorting result; and determine at least one candidate resource ranked first in the sorting result as a resource to be recommended.

[0074] As an alternative, the electronic device may determine, for each candidate resource in the candidate resource set, whether the predicted click-through rate of the candidate resource is greater than or equal to a predicted click-through rate threshold; if the predicted click-through rate of the candidate resource is greater than or equal to the predicted click-through rate threshold, the candidate resource is determined as a resource to be recommended.

[0075] Among them, determining the candidate resources with a larger predicted click rate as the resources to be recommended can increase the probability of being clicked on the resources to be recommended and reduce the number of times the target object refreshes the resources to be recommended, thereby further improving the efficiency of resource recommendation.

[0076] In another example, the process of the electronic device executing step 305 may, for example, be to determine the value of a candidate resource in the candidate resource set on at least one recommendation indicator; determine a score of the candidate resource based on the value of the candidate resource on at least one recommendation indicator and the predicted click-through rate; sort the candidate resources in the candidate resource set in descending order based on the score to obtain a sorting result; and determine at least one candidate resource ranked first in the sorting result as a resource to be recommended.

[0077] The recommendation index may include at least one of the following: browsing time index, interaction index, browsing index, etc., which can be set according to actual needs. The interaction index may include at least one of the following: collection index, comment index, like index, share index, subscription index, download index, etc.

[0078] Among them, the process of the electronic device determining the score of the candidate resource can, for example, be that the electronic device determines the weight of at least one recommendation indicator and the predicted click-through rate; and performs weighted summation processing on the value of at least one recommendation indicator and the predicted click-through rate according to the weight to obtain the score of the candidate resource.

[0079] Among them, combining the numerical value of at least one recommendation indicator and the predicted click-through rate to determine the score of the candidate resource can improve the accuracy of the determined score, thereby further improving the match between the recommended resource and the content of interest to the target object, and further improving the accuracy of resource recommendation.

[0080] Step 306: Perform resource recommendation processing on the target object according to the resources to be recommended.

[0081] It should be noted that the details of steps 301 to 302 and step 306 can be found in Figure 1 Steps 101 to 102 and step 104 in the illustrated embodiment will not be described in detail herein.

[0082] The resource recommendation method of the embodiment of the present disclosure obtains a candidate resource set and a historical resource sequence of the target object in a historical time period before the current time point; determines an interest representation vector of the target object at the current time point based on the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point; determines the representation vectors of the candidate resources in the candidate resource set; determines the predicted click-through rate of the candidate resources based on the interest representation vector and the representation vectors of the candidate resources; selects the resources to be recommended from the candidate resource set based on the predicted click-through rate; performs resource recommendation processing on the target object based on the resources to be recommended; wherein, the predicted click-through rate of the candidate resources is determined based on the interest representation vector and the representation vectors of the candidate resources in the candidate resource set, and then the resources to be recommended are selected based on the predicted click-through rate, which can further improve the matching degree between the determined resources to be recommended and the content of interest to the target object, and further improve the accuracy of resource recommendation; and accurate recommendation of resources can reduce the number of resource recommendations, thereby improving the efficiency of resource recommendation.

[0083] Figure 4 It is a schematic diagram according to the fourth embodiment of the present disclosure. It should be noted that the training method of the resource recommendation model of the embodiment of the present disclosure can be applied to the training device of the resource recommendation model. The device can be configured in an electronic device so that the electronic device can perform the training function of the resource recommendation model.

[0084] Among them, the electronic device can be any device with computing capabilities, such as a personal computer (PC), a mobile terminal, a server, etc. The mobile terminal can be, for example, a vehicle-mounted device, a mobile phone, a tablet computer, a personal digital assistant, a wearable device, a smart speaker, a server, a server cluster, and other hardware devices with various operating systems, touch screens and / or display screens.

[0085] The training device of the resource recommendation model may also be software in an electronic device, such as training software of the resource recommendation model, etc. In the following embodiments, the execution subject is an electronic device as an example for description.

[0086] like Figure 4 As shown, the training method of the resource recommendation model may include the following steps:

[0087] Step 401, obtaining training data; the training data includes structured content of sample historical resources in a sample historical resource sequence, structured content of sample recommendation resources, and sample click resources in sample recommendation resources.

[0088] In the embodiment of the present disclosure, the sample history resource may be a sample recommendation resource for a sample object, a sample object click history resource, and a sample click resource may be a resource clicked by a sample object in the sample recommendation resource.

[0089] The structured content of the sample history resource may include at least one metadata of the sample history resource. The metadata in the structured content may include at least one of the following: the title of the resource, keywords in the resource, the field to which the resource belongs, and the publishing object of the resource.

[0090] Step 402, obtaining an initial resource recommendation model; the resource recommendation model includes an encoding network, a decoding network, and a fully connected network that are connected in sequence.

[0091] In the disclosed embodiment, the encoding network is used to determine the representation vector of the sample historical resource in combination with the structured content of the sample historical resource, and to determine the representation vector of the sample recommendation resource in combination with the structured content of the sample recommendation resource. The decoding network is used to determine the interest representation vector of the sample object to which the sample historical resource belongs in combination with the representation vector of the sample historical resource in the sample historical resource sequence. The fully connected network is used to determine the predicted click rate of the sample recommendation resource in combination with the interest representation vector and the representation vector of the sample recommendation resource.

[0092] Among them, a multi-head attention mechanism can be set in the encoding network to learn the association between various metadata in the structured content of the resource, so as to extract the deep features in the structured content and obtain the representation vector of the resource.

[0093] Among them, a multi-head attention mechanism can also be set in the decoding network to learn the correlation between the representation vectors of each sample historical resource, learn the change trend of the resources of interest to the target object, and thus determine the interest representation vector of the target object.

[0094] Step 403 , the initial resource recommendation model is trained using the structured content of the sample historical resources, the structured content of the sample recommended resources, and the sample click resources to obtain a trained resource recommendation model.

[0095] In the embodiment of the present disclosure, the process of the electronic device executing step 403 may, for example, be as follows: inputting the structured content of the sample historical resources and the structured content of the sample recommended resources into the encoding network in the resource recommendation model respectively to obtain the representation vector of the sample historical resources and the representation vector of the sample recommended resources; inputting the representation vector of each sample historical resource in the sample historical resource sequence into the decoding network in the resource recommendation model to obtain the interest representation vector of the object to which the sample historical resource belongs; inputting the interest representation vector and the representation vector of the sample recommended resource into the fully connected network in the resource recommendation model to obtain the predicted click probability of the sample recommended resource; determining the loss function value by combining the sample click resources in the sample recommended resources, the predicted click probability of the sample recommended resources, and the loss function of the resource recommendation model; and adjusting the parameters of the encoding network, the decoding network, and the fully connected network by combining the loss function value to obtain the trained resource recommendation model.

[0096] The electronic device combines the sample click resources in the sample recommendation resources, the predicted click probability of the sample recommendation resources, and the loss function of the resource recommendation model to determine the value of the loss function. For example, for the sample recommendation resources, the actual click probability of the sample recommendation resources can be determined in combination with whether the sample recommendation resources are sample click resources; the value of the loss function can be determined in combination with the predicted click probability, the actual click probability and the loss function of the sample recommendation resources. The loss function is used to calculate the difference between the predicted click probability and the actual click probability. The smaller the difference, the smaller the value of the loss function; the larger the difference, the larger the value of the loss function.

[0097] Among them, when the sample recommendation resource is a sample click resource, it can be determined that the actual click probability of the sample recommendation resource is 1; when the sample recommendation resource is not a sample click resource, it can be determined that the actual click probability of the sample recommendation resource is 0.

[0098] Among them, the predicted click probability of the sample recommended resources is determined by combining the structured content of each sample historical resource in the sample historical resource sequence and the structured content of the sample recommended resources; the resource recommendation model is parameter adjusted by combining the predicted click probability and the actual click probability of the sample recommended resources, so that the predicted click probability of the sample recommended resources determined by the trained resource recommendation model can tend to the actual click probability of the sample recommended resources, thereby improving the accuracy of the trained resource recommendation model.

[0099] The training method of the resource recommendation model of the embodiment of the present disclosure is through obtaining training data; the training data includes structured content of sample historical resources in a sample historical resource sequence, structured content of sample recommended resources, and sample click resources in the sample recommended resources; obtaining an initial resource recommendation model; the resource recommendation model includes an encoding network, a decoding network, and a fully connected network connected in sequence; the initial resource recommendation model is trained using the structured content of sample historical resources, the structured content of sample recommended resources, and the sample click resources to obtain a trained resource recommendation model; wherein, the initial resource recommendation model is trained using the structured content of sample historical resources, the structured content of sample recommended resources, and the sample click resources, so that the resource recommendation model can learn the association relationship between each sample historical resource and the interest representation vector of the learning object, thereby improving the accuracy of the trained resource recommendation model.

[0100] In order to implement the above embodiment, the present disclosure also provides a resource recommendation device. Figure 5 As shown, Figure 5 The fifth embodiment of the present disclosure is shown in FIG. 5. The resource recommendation device 50 may include: an acquisition module 501 , a determination module 502 , a selection module 503 and a recommendation processing module 504 .

[0101] Among them, the acquisition module 501 is used to obtain a candidate resource set and a historical resource sequence of the target object in a historical time period before the current time point; the determination module 502 is used to determine the interest representation vector of the target object at the current time point based on the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point; the selection module 503 is used to select the resources to be recommended from the candidate resource set according to the interest representation vector; the recommendation processing module 504 is used to perform resource recommendation processing on the target object based on the resources to be recommended.

[0102] As a possible implementation method of an embodiment of the present disclosure, the determination module 502 includes a first determination unit and a second determination unit; the first determination unit is used to determine the representation vector of the historical resources in the historical resource sequence; the second determination unit is used to determine the interest representation vector of the target object at the current time point based on the representation vector of each historical resource in the historical resource sequence.

[0103] As a possible implementation method of an embodiment of the present disclosure, the first determination unit is specifically used to determine the structured content of the historical resources in the historical resource sequence; the structured content includes at least one metadata in the historical resources; the structured content is input into the encoding network in the resource recommendation model, and the representation vector of the historical resources output by the encoding network is obtained.

[0104] As a possible implementation of the embodiment of the present disclosure, the metadata in the structured content includes at least one of the following: a title of the resource, keywords in the resource, a field to which the resource belongs, and a publishing object of the resource.

[0105] As a possible implementation method of an embodiment of the present disclosure, the second determination unit is specifically used to determine a representation vector sequence based on the representation vectors of each historical resource in the historical resource sequence; input the representation vector sequence into a decoding network in a resource recommendation model, and obtain the interest representation vector output by the decoding network.

[0106] As a possible implementation method of the embodiment of the present disclosure, the determination module 502 also includes: a division processing unit, a third determination unit, a splicing processing unit and an update processing unit; the division processing unit is used to divide and process each historical resource in the historical resource sequence according to the collection time point to obtain multiple historical resource combinations; the third determination unit is used to determine the combination number of the historical resource combination and the number vector corresponding to the combination number; the splicing processing unit is used to, for each historical resource in the historical resource sequence, splice the number vector corresponding to the combination number of the historical resource combination to which the historical resource belongs with the representation vector of the historical resource to obtain a processed representation vector; the update processing unit is used to update the representation vector of the historical resource according to the processed representation vector.

[0107] As a possible implementation method of the embodiment of the present disclosure, the acquisition module 501 is specifically used to perform resource recall processing and resource screening processing according to the attribute information of the target object to obtain the candidate resource set; obtain the historical resources clicked by the target object in the historical time period before the current time point, and the click time point of the historical resources; sort each of the historical resources according to the click time point to obtain the historical resource sequence.

[0108] As a possible implementation method of the embodiment of the present disclosure, the acquisition module 501 is specifically used to perform resource recall processing according to the attribute information of the target object to obtain a recalled resource set; determine the first click rate of the recalled resources in the recalled resource set based on the attribute information and the first click rate model; filter the recalled resources in the recalled resource set based on the first click rate to obtain a filtered resource set; determine the second click rate of the filtered resources in the filtered resource set based on the attribute information and the second click rate model; the parameter amount of the second click rate model is greater than the parameter amount of the first click rate model; filter the filtered resources in the filtered resource set based on the second click rate to obtain the candidate resource set.

[0109] As a possible implementation method of an embodiment of the present disclosure, the selection module 503 includes a fourth determination unit, a fifth determination unit and a selection unit; the fourth determination unit is used to determine the representation vector of the candidate resource in the candidate resource set; the fifth determination unit is used to determine the predicted click-through rate of the candidate resource based on the interest representation vector and the representation vector of the candidate resource; the selection unit is used to select the resource to be recommended from the candidate resource set based on the predicted click-through rate.

[0110] As a possible implementation manner of an embodiment of the present disclosure, the fifth determination unit is specifically used to input the interest representation vector and the representation vector of the candidate resource into a fully connected network in a resource recommendation model, and obtain the predicted click rate output by the fully connected network.

[0111] As a possible implementation method of an embodiment of the present disclosure, the selection unit is specifically used to sort the candidate resources in the candidate resource set in descending order according to the predicted click-through rate to obtain a sorting result; and determine at least one candidate resource ranked first in the sorting result as the resource to be recommended.

[0112] As a possible implementation method of the embodiment of the present disclosure, the selection unit is specifically used to determine the value of the candidate resource in the candidate resource set on at least one recommendation indicator; determine the score of the candidate resource based on the value of the candidate resource on at least one recommendation indicator and the predicted click-through rate; sort the candidate resources in the candidate resource set in descending order based on the score to obtain a sorting result; and determine at least one candidate resource ranked first in the sorting result as the resource to be recommended.

[0113] The resource recommendation device of the embodiment of the present disclosure obtains a candidate resource set and a historical resource sequence of the target object in a historical time period before the current time point; determines an interest representation vector of the target object at the current time point based on the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point; selects a to-be-recommended resource from the candidate resource set based on the interest representation vector; and performs resource recommendation processing on the target object based on the to-be-recommended resource; wherein the interest representation vector determined in combination with the historical resource sequence can represent the resources that the target object is interested in at the current time point, so that the determined to-be-recommended resource is the resource that the target object currently wants to click, thereby improving the matching degree between the to-be-recommended resource and the content of interest to the target object, and improving the accuracy of resource recommendation; and accurate recommendation of resources can reduce the number of resource recommendations, thereby improving resource recommendation efficiency.

[0114] In order to implement the above embodiment, the present disclosure also provides a training device for a resource recommendation model. Figure 6 As shown, Figure 6 6 is a schematic diagram according to the sixth embodiment of the present disclosure. The resource recommendation model training device 60 may include: a first acquisition module 601 , a second acquisition module 602 and a training processing module 603 .

[0115] Among them, the first acquisition module 601 is used to obtain training data; the training data includes the structured content of sample historical resources in the sample historical resource sequence, the structured content of sample recommendation resources, and the sample click resources in the sample recommendation resources; the second acquisition module 602 is used to obtain the initial resource recommendation model; the resource recommendation model includes an encoding network, a decoding network and a fully connected network connected in sequence; the training processing module 603 is used to use the structured content of the sample historical resources, the structured content of the sample recommendation resources and the sample click resources to train the initial resource recommendation model to obtain a trained resource recommendation model.

[0116] As a possible implementation method of the embodiment of the present disclosure, the training processing module 603 is specifically used to input the structured content of the sample historical resources and the structured content of the sample recommended resources into the encoding network in the resource recommendation model respectively to obtain the representation vector of the sample historical resources and the representation vector of the sample recommended resources; input the representation vector of each sample historical resource in the sample historical resource sequence into the decoding network in the resource recommendation model to obtain the interest representation vector of the object to which the sample historical resource belongs; input the interest representation vector and the representation vector of the sample recommended resource into the fully connected network in the resource recommendation model to obtain the predicted click probability of the sample recommended resource; determine the loss function value in combination with the sample click resources in the sample recommended resources, the predicted click probability of the sample recommended resources, and the loss function of the resource recommendation model; and perform parameter adjustment processing on the encoding network, the decoding network, and the fully connected network in combination with the loss function value to obtain a trained resource recommendation model.

[0117] The training device of the resource recommendation model of the embodiment of the present disclosure obtains training data; the training data includes structured content of sample historical resources in a sample historical resource sequence, structured content of sample recommended resources, and sample click resources in the sample recommended resources; obtains an initial resource recommendation model; the resource recommendation model includes an encoding network, a decoding network, and a fully connected network connected in sequence; the initial resource recommendation model is trained using the structured content of sample historical resources, the structured content of sample recommended resources, and the sample click resources to obtain a trained resource recommendation model; wherein, the initial resource recommendation model is trained using the structured content of sample historical resources, the structured content of sample recommended resources, and the sample click resources, so that the resource recommendation model can learn the association relationship between each sample historical resource and the interest representation vector of the learning object, thereby improving the accuracy of the trained resource recommendation model.

[0118] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information are all carried out with the user's consent, comply with the relevant laws and regulations, and do not violate public order and good morals.

[0119] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0120] Figure 7 A schematic block diagram of an example electronic device 700 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0121] like Figure 7 As shown, the device 700 includes a computing unit 701, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the device 700 can also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0122] A number of components in the device 700 are connected to the I / O interface 705, including: an input unit 706, such as a keyboard, a mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a disk, an optical disk, etc.; and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the device 700 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0123] The computing unit 701 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 701 performs the various methods and processes described above, such as a resource recommendation method or a training method for a resource recommendation model. For example, in some embodiments, the resource recommendation method or the training method for a resource recommendation model may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 708. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 700 via the ROM 702 and / or the communication unit 709. When the computer program is loaded into the RAM 703 and executed by the computing unit 701, one or more steps of the resource recommendation method or the training method for the resource recommendation model described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to execute the resource recommendation method or the resource recommendation model training method in any other appropriate manner (for example, by means of firmware).

[0124] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0125] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0126] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0127] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0128] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0129] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0130] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0131] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A resource recommendation method, the method comprising: Obtain a candidate resource set and a historical resource sequence of the target object in the historical time period before the current time point; Determine, according to the historical resource sequence, an interest representation vector of the target object at the current time point; the interest representation vector represents the resources that the target object is interested in at the current time point; Selecting a resource to be recommended from the candidate resource set according to the interest representation vector; According to the resources to be recommended, resource recommendation processing is performed on the target object.

2. The method according to claim 1, wherein: The determining, according to the historical resource sequence, the interest representation vector of the target object at the current time point includes: Determining a representation vector of a historical resource in the historical resource sequence; The interest representation vector of the target object at the current time point is determined according to the representation vector of each historical resource in the historical resource sequence.

3. The method according to claim 2, wherein: The determining of the representation vector of the historical resource in the historical resource sequence includes: Determining structured content of historical resources in the historical resource sequence; the structured content includes at least one metadata of the historical resources; The structured content is input into an encoding network in a resource recommendation model, and a representation vector of the historical resource output by the encoding network is obtained.

4. The method according to claim 3, wherein: The metadata in the structured content includes at least one of the following: a title of the resource, a keyword in the resource, a field to which the resource belongs, and a publishing object of the resource.

5. The method according to claim 2, wherein: The step of determining the interest representation vector of the target object at the current time point according to the representation vector of each historical resource in the historical resource sequence includes: Determining a characterization vector sequence according to the characterization vector of each historical resource in the historical resource sequence; The representation vector sequence is input into a decoding network in a resource recommendation model to obtain the interest representation vector output by the decoding network.

6. The method according to claim 2 or 5, wherein: After determining the characterization vector of the historical resource in the historical resource sequence, the method further includes: Each historical resource in the historical resource sequence is divided and processed according to the collection time point to obtain multiple historical resource combinations; Determine a combination number of the historical resource combination and a number vector corresponding to the combination number; For each historical resource in the historical resource sequence, concatenate the number vector corresponding to the combination number of the historical resource combination to which the historical resource belongs with the representation vector of the historical resource to obtain a processed representation vector; The representation vector of the historical resource is updated according to the processed representation vector.

7. The method according to claim 1, wherein: The step of obtaining the candidate resource set and the historical resource sequence of the target object in the historical time period before the current time point includes: Perform resource recall processing and resource screening processing according to the attribute information of the target object to obtain the candidate resource set; Obtain the historical resources clicked by the target object in the historical time period before the current time point, as well as the click time points of the historical resources; The historical resources are sorted according to the click time points to obtain the historical resource sequence.

8. The method according to claim 7, wherein: The performing resource recall processing and resource screening processing according to the attribute information of the target object to obtain the candidate resource set includes: Perform resource recall processing according to the attribute information of the target object to obtain a recalled resource set; Determine the first click rate of the recalled resource in the recalled resource set by combining the attribute information and the first click rate model; According to the first click rate, the recalled resources in the recalled resource set are screened to obtain a screened resource set; Determine a second click rate of the screening resource in the screening resource set in combination with the attribute information and a second click rate model; the parameter amount of the second click rate model is greater than the parameter amount of the first click rate model; According to the second click rate, screening resources in the screening resource set are screened to obtain the candidate resource set.

9. The method according to claim 1, wherein: The selecting a resource to be recommended from the candidate resource set according to the interest representation vector includes: Determining a representation vector of a candidate resource in the candidate resource set; Determining a predicted click rate of the candidate resource according to the interest representation vector and the representation vector of the candidate resource; The resource to be recommended is selected from the candidate resource set according to the predicted click rate.

10. The method according to claim 9, wherein: The determining, according to the interest representation vector and the representation vector of the candidate resource, the predicted click rate of the candidate resource includes: The interest representation vector and the representation vector of the candidate resource are input into a fully connected network in a resource recommendation model, and the predicted click rate output by the fully connected network is obtained.

11. The method according to claim 9, wherein: The selecting the resource to be recommended from the candidate resource set according to the predicted click rate includes: According to the predicted click-through rate, the candidate resources in the candidate resource set are sorted in descending order to obtain a sorting result; At least one candidate resource ranked first in the ranking result is determined as the resource to be recommended.

12. The method according to claim 9, wherein: The selecting the resource to be recommended from the candidate resource set according to the predicted click rate includes: Determine a value of the candidate resource in the candidate resource set on at least one recommendation indicator; Determining a score of the candidate resource according to a value of the candidate resource on at least one recommendation indicator and the predicted click-through rate; According to the scores, the candidate resources in the candidate resource set are sorted in descending order to obtain a sorting result; At least one candidate resource ranked first in the ranking result is determined as the resource to be recommended.

13. A method for training a resource recommendation model, the method comprising: Get training data; The training data includes structured content of sample historical resources in a sample historical resource sequence, structured content of sample recommendation resources, and sample click resources in the sample recommendation resources; Acquire an initial resource recommendation model; the resource recommendation model includes an encoding network, a decoding network, and a fully connected network connected in sequence; The initial resource recommendation model is trained using the structured content of the sample historical resources, the structured content of the sample recommended resources, and the sample click resources to obtain a trained resource recommendation model.

14. The method according to claim 13, wherein: The training process of the initial resource recommendation model using the structured content of the sample history resource, the structured content of the sample recommendation resource, and the structured content of the sample click resource to obtain a trained resource recommendation model includes: Inputting the structured content of the sample historical resource and the structured content of the sample recommended resource into the encoding network in the resource recommendation model respectively, and obtaining the representation vector of the sample historical resource and the representation vector of the sample recommended resource; Inputting the representation vector of each sample historical resource in the sample historical resource sequence into a decoding network in the resource recommendation model to obtain an interest representation vector of the object to which the sample historical resource belongs; Inputting the interest representation vector and the representation vector of the sample recommended resource into a fully connected network in the resource recommendation model to obtain a predicted click probability of the sample recommended resource; Determine a loss function value by combining the sample click resources in the sample recommendation resources, the predicted click probability of the sample recommendation resources, and the loss function of the resource recommendation model; In combination with the loss function value, parameter adjustment processing is performed on the encoding network, the decoding network, and the fully connected network to obtain a trained resource recommendation model.

15. A resource recommendation device, comprising: An acquisition module is used to acquire a candidate resource set and a historical resource sequence of a target object in a historical time period before a current time point; A determination module, configured to determine an interest representation vector of the target object at the current time point according to the historical resource sequence; the interest representation vector represents the resources that the target object is interested in at the current time point; A selection module, configured to select a resource to be recommended from the candidate resource set according to the interest representation vector; The recommendation processing module is used to perform resource recommendation processing on the target object according to the resources to be recommended.

16. The device according to claim 15, wherein: The determination module includes a first determination unit and a second determination unit; The first determining unit is used to determine the representation vector of the historical resource in the historical resource sequence; The second determining unit is configured to determine the interest representation vector of the target object at the current time point according to the representation vector of each historical resource in the historical resource sequence.

17. The device according to claim 16, wherein: The first determining unit is specifically configured to: Determining structured content of historical resources in the historical resource sequence; the structured content includes at least one metadata of the historical resources; The structured content is input into an encoding network in a resource recommendation model, and a representation vector of the historical resource output by the encoding network is obtained.

18. The device according to claim 17, wherein: The metadata in the structured content includes at least one of the following: a title of the resource, a keyword in the resource, a field to which the resource belongs, and a publishing object of the resource.

19. The device according to claim 16, wherein: The second determining unit is specifically configured to: Determining a characterization vector sequence according to the characterization vector of each historical resource in the historical resource sequence; The representation vector sequence is input into a decoding network in a resource recommendation model to obtain the interest representation vector output by the decoding network.

20. The device according to claim 16 or 19, wherein: The determination module further includes: a division processing unit, a third determination unit, a splicing processing unit and an update processing unit; The division processing unit is used to divide and process each historical resource in the historical resource sequence according to the collection time point to obtain multiple historical resource combinations; The third determining unit is used to determine the combination number of the historical resource combination and the number vector corresponding to the combination number; The splicing processing unit is used to perform splicing processing on the number vector corresponding to the combination number of the historical resource combination to which the historical resource belongs and the representation vector of the historical resource for each historical resource in the historical resource sequence, so as to obtain a processed representation vector; The update processing unit is used to update the representation vector of the historical resource according to the processed representation vector.

21. The device according to claim 15, wherein: The acquisition module is specifically used for: Perform resource recall processing and resource screening processing according to the attribute information of the target object to obtain the candidate resource set; Obtain the historical resources clicked by the target object in the historical time period before the current time point, as well as the click time points of the historical resources; The historical resources are sorted according to the click time points to obtain the historical resource sequence.

22. The device according to claim 21, wherein The acquisition module is also specifically used for: Perform resource recall processing according to the attribute information of the target object to obtain a recalled resource set; Determine the first click rate of the recalled resource in the recalled resource set by combining the attribute information and the first click rate model; According to the first click rate, the recalled resources in the recalled resource set are screened to obtain a screened resource set; Determine a second click rate of the screening resources in the screening resource set by combining the attribute information and the second click rate model; The parameter amount of the second click rate model is greater than the parameter amount of the first click rate model; According to the second click rate, screening resources in the screening resource set are screened to obtain the candidate resource set.

23. The device according to claim 15, wherein: The selection module includes a fourth determination unit, a fifth determination unit and a selection unit; The fourth determining unit is used to determine the representation vector of the candidate resource in the candidate resource set; The fifth determining unit is used to determine the predicted click rate of the candidate resource according to the interest representation vector and the representation vector of the candidate resource; The selection unit is used to select the resource to be recommended from the candidate resource set according to the predicted click rate.

24. The device according to claim 23, wherein: The fifth determining unit is specifically configured to: The interest representation vector and the representation vector of the candidate resource are input into a fully connected network in a resource recommendation model, and the predicted click rate output by the fully connected network is obtained.

25. The device according to claim 23, wherein: The selection unit is specifically used to: According to the predicted click-through rate, the candidate resources in the candidate resource set are sorted in descending order to obtain a sorting result; At least one candidate resource ranked first in the ranking result is determined as the resource to be recommended.

26. The device according to claim 23, wherein The selection unit is specifically used to: Determine a value of the candidate resource in the candidate resource set on at least one recommendation indicator; Determining a score of the candidate resource according to a value of the candidate resource on at least one recommendation indicator and the predicted click-through rate; According to the scores, the candidate resources in the candidate resource set are sorted in descending order to obtain a sorting result; At least one candidate resource ranked first in the ranking result is determined as the resource to be recommended.

27. A training device for a resource recommendation model, the device comprising: A first acquisition module, used to acquire training data; The training data includes structured content of sample historical resources in a sample historical resource sequence, structured content of sample recommendation resources, and sample click resources in the sample recommendation resources; A second acquisition module is used to acquire an initial resource recommendation model; the resource recommendation model includes an encoding network, a decoding network, and a fully connected network connected in sequence; The training processing module is used to use the structured content of the sample historical resources, the structured content of the sample recommendation resources and the sample click resources to train the initial resource recommendation model to obtain a trained resource recommendation model.

28. The device according to claim 27, wherein The training processing module is specifically used to: Inputting the structured content of the sample historical resource and the structured content of the sample recommended resource into the encoding network in the resource recommendation model respectively, and obtaining the representation vector of the sample historical resource and the representation vector of the sample recommended resource; Inputting the representation vector of each sample historical resource in the sample historical resource sequence into a decoding network in the resource recommendation model to obtain an interest representation vector of the object to which the sample historical resource belongs; Inputting the interest representation vector and the representation vector of the sample recommended resource into a fully connected network in the resource recommendation model to obtain a predicted click probability of the sample recommended resource; Determine a loss function value by combining the sample click resources in the sample recommendation resources, the predicted click probability of the sample recommendation resources, and the loss function of the resource recommendation model; In combination with the loss function value, parameter adjustment processing is performed on the encoding network, the decoding network, and the fully connected network to obtain a trained resource recommendation model.

29. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 12; Alternatively, the method according to any one of claims 13 to 14 is performed.

30. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 12; or to execute the method according to any one of claims 13 to 14.

31. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the method according to any one of claims 1 to 12; or implements the method according to any one of claims 13 to 14.

Images